中文摘要：

针对异构多核环境下的任务调度问题,为充分发挥异构多核平台优势,提出一种基于改进的混洗蛙跳算法的异构多核处理器任务调度算法——ISFLA算法。ISFLA算法通过建立合适的个体编码方案和位置更新方法实现个体搜索空间到离散空间的映射,通过调整子群更新方法规避在离散空间下标准算法中子群更新易于停滞在最差值的情况,使得标准SFLA算法可以应用于异构多核环境下的任务调度问题。同时算法引入了遗传算法的交叉操作,通过将全局较优解群体的信息扩散到种群,增加了算法跳出局部极优值的能力,优化了算法的全局搜索能力。最后通过实验证明,对比常见的遗传算法,ISFLA算法具有更好的求解能力,可以提供更少执行时间的任务调度方案,具有较好的应用价值。

英文摘要：

For the high performance in muhi-core processor environment, this paper proposed a heterogeneous multi-core task scheduling algorithm, which was based on an improved shuffled frog-leaping algorithm named ISFLA. This ISFLA improved the local interaction and the global exchange of information based on the basic SFLA, avoided the local interact operation stucking at the local worse individual and increased the ability of the global searching. The result of comparing with the GA algorithm in five experiments demonstrates that the ISFLA is better than GA. ISFLA is suitable for the heterogeneous multi-core processor.